Basic functions of a computer include information processing and storage. In typical computer systems, these arithmetic, logic, and memory operations are performed by devices that are capable of reversibly switching between two states often referred to as “0” and “1.” In most cases, such switching devices are fabricated from semiconducting devices that perform these various functions and are capable of switching between two states at a very high speed using minimum amounts of electrical energy. Thus, for example, transistors and transistor variants perform the basic switching and storage functions in computers.
Because of the huge data storage requirements of modem computers, a new, compact, low-cost, very high capacity, high speed memory configuration is needed. To reach this objective, molecular electronic switches, wires, microsensors for chemical analysis, and opto-electronic components for use in optical computing have been pursued. The principal advantages of using molecules in these applications are high component density (upwards of 1018 bits per square centimeter), increased response speeds, and high energy efficiency.
A variety of approaches have been proposed for molecular-based memory devices. While these approaches generally employ molecular architectures that can be switched between two different states, all of the approaches described to date have intrinsic limitations making their uses in computational devices difficult or impractical.
For example, such approaches to the production of molecular memories have involved photochromic dyes, electrochromic dyes, redox dyes, and molecular machines. Each of these approaches, however, has intrinsic limitations that ultimately render it unsuitable for use in molecular memories. For example, photochromic dyes change conformation in response to the absorption of light (e.g. cis-trans interconversion of an alkene, ring opening of a spiropyran, interconversion between excited-states in bacteriorhodopsin, etc.). Typically, the molecular structure of the dye is interconverted between two states that have distinct spectral properties.
Reading and writing data with such photochromic dyes requires use of light, often in the visible region (400–700 nm). Light-mediated data storage has intrinsic diffraction-limited size constraints. Moreover, most photochromic schemes are limited to scanning and interrogating dyes deposited on a surface and are not amenable to 3-D data storage. Even with near-field optical approaches, which might allow reliable encoding/reading of data elements of 100×100 nm dimensions (Nieto-Vesperinas and Garcia, N., eds. (1996) Optics at the Nanometer Scale, NATO ASI Series E, Vol. 319, Kluwer Academic Publishers: Dordrecht) the inherent restricted dimensionality (2-D) limits data density to 1010 bits/cm2. Strategies for 3-dimensional reading and writing of photochromic systems have been proposed that rely on two-photon excitation of dyes to encode data, and one-photon excitation to read the data (Birge et al. (1994) Amer. Sci. 82: 349–355, Parthenopoulos and Rentzepis (1989) Science, 245: 843–845), but it is believed that no high-density memory cubes have reached prototype stage in spite of the passage of at least a decade since their initial proposition. In addition, it is noted that these dyes often exhibit relatively slow switching times ranging from microsecond to millisecond durations.
Electrochromic dyes have been developed that undergo a slight change in absorption spectrum upon application of an applied electric field (Liptay (1969) Angew. Chem., Int. Ed. Engi. 8: 177–188). The dyes must be oriented in a fixed direction with respect to the applied field. Quite high fields (>107 V/cm) must be applied to observe an altered absorption spectrum which can result in heat/power dissipation problems. In addition, the change in the absorption spectrum is typically quite small, which can present detection difficulties. The dyes revert to the initial state when the applied field is turned off.
Redox dyes have been developed that undergo a change in absorption spectrum upon chemical or electrochemical reduction (typically a 2-electron, 2-proton reduction) (Otsuki et al. (1996) Chem. Lett. 847–848). Such systems afford bistable states (e.g., quinone/hydroquinone, azo/hydrazo). Redox dyes have only been examined in solution studies, where they have been proposed for applications as switches and sensors (de Silva et al. (1997) Chem. Rev. 97: 1515–1566). On a solid substrate, electrochemical reduction would need to be accompanied by a source of protons. The latter requirement may be difficult to achieve on a solid substrate. Furthermore, any optical reading scheme would pose the same 2-D limitations as described for photochromic dyes.
Yet another approach involves the design of molecular machines (Anell et al. (1992) J. Am. Chem. Soc. 114: 193–218). These elegant molecular architectures have moving parts that can be switched from one position to another by chemical or photochemical means. The chemically induced systems have applications as sensors but are not practical for memory storage, while the photochemically induced systems have the same fundamental limitations as photochromic dyes. Moreover, methods have not yet been developed for delineating the conformation/structure of the molecular machine that are practical in any device applications. 1H NMR spectroscopy, for example, is clearly the method of choice for elucidating structure/conformation for molecules in solution, but is totally impractical for interrogating a molecular memory element. None of the current architectures for molecular machines has been designed for assembly on a solid substrate, an essential requirement in a viable device.
In summary, photochromic dyes, electrochromic dyes, redox-sensitive dyes, and molecular machines all have fundamental limitations that have precluded their application as viable memory elements. These molecular architectures are typically limited by reading/writing constraints. Furthermore, even in cases where the effective molecular bistability is obtained, the requirement for photochemical reading restricts the device architecture to a 2-dimensional thin film. The achievable memory density of such a film is unlikely to exceed 1010 bits/cm2. Such limitations greatly diminish the appeal of these devices as viable molecular memory elements.